C7+ Characterization
This section will be utilizing examples from the projects specified in this section. Make sure to add these projects first, if you want to follow along with the examples.
EOS Type and Molecular Weight Model
What would we recommend if this is your first time?
Since the choice of EOS model type shouldn't impact the final predictions, we recommend sticking with the default choice.
If you have a limited number of samples, we recommend using the molecular weights from the correlation instead of the gamma model values. We also don't recommend using LMW regression options, unless you have a large range of average molecular weights for your flashed oil fluids!
If you want to see the impact of the choice of molecular weight correlation, you can save a case for each correlation type and compare the results.
There are two EOS models that are available in whitsonPVT when developing a new fluid model. These are the Peng-Robinson (PR) and Soave-Relich-Kwong (SRK) models. There is no fundamental or significant difference in the ability of these two models to predict hydrocarbon phase behavior, so the choice between these two models is one of preference.
The molecular weight model choices is between using the gamma model or correlation based single carbon number (SCN) component molecular weights for your C7+ components.
The molecular weight (MW) correlation options are the Effective Paraffin model and the Twu model. Even if the gamma model is chosen to use for estimating the component molecular weights, the correlation type is still used to initialize the lower molecular weights (LMW) in the gamma model. If LMW tuning is applied and the gamma model is selected for the MW model, the correlation type becomes irrelevant!
Gamma Model
You can see the results from the gamma model calculation with default parameters in whitsonPVT, where the gamma model estimates component mass amounts for the various examples below.
Example 1 - Sample CL-63169 (Duvernay) - Single Gas Sample
Example 2 - Sample CL-70055 (Duvernay) - Single Oil Sample
Example 3 - Field-Wide Fluid Model (Volve Field)
Example 3 shows how a field-wide EOS model for the volve field predicts the composition for a single sample (6103-MA) within the Volve field in Norway. The measured point label now specifies "(Decontaminated)" as the sample selected has oil-based mud contamination associated with it.
Example 4 - Basin-Wide Fluid Model (Duvernay Basin)
As with the previous example, we here see how a field-wide EOS model has been used to predict the composition of sample 'CL-77279'.
Specific Gravity Model
The Specific Gravity Model can be used to estimate SG and flashed oil densities. The deviation between predicted and measured values for the flashed oil densities are shown below. Note that for the single-sample calculations, there won't be much of a deviation, as the model has only been used to match a single data point.
Example 1 - Sample CL-63169 (Duvernay) - Single Gas Sample
Example 2 - Sample CL-70055 (Duvernay) - Single Oil Sample
Example 3 - Field-Wide Fluid Model (Volve Field)
Example 4 - Basin-Wide Fluid Model (Duvernay Basin)
Results and Saved Cases
Once you have run a regression where you are satisfied with the results, it is also important to save your results as a 'Case'. You can do this by clicking the 'Save Case' button marked in Red on the picture below. You will then need to select a name for your new case, which will be stored in the 'Cases' table marked in Yellow.
The 'Current' case represents whatever parameters you used for your last regression, and will be equal to your last saved case if you have not re-ran a regression after having saved it. If you want to compare the results of different cases, you can tick the box next to your saved cases (marked in Red below), and the results will overlap with your 'current' case results.
The following pictures illustrate how the Twu MW correlation type (Field-Wide_Standard case) compares to the Effective Paraffin correlation (Current case) in the Volve full-field fluid model (Example 3).
Compositional Estimation:
FLO Density Estimation:
Next Steps
The next step in the fluid model development is to tune the initial characterized EOS model!